“I would like to understand the extent to which we understand things,” the MIT economist says.
Peter Dizikes | MIT News
When you read about a new study, you may wonder: How accurate are these results? MIT economist Isaiah Andrews PhD ’14 often asks that as well, especially about social sciences research. Unlike most of us, though, Andrews’ job involves answering that question.
Andrews, a professor in MIT’s Department of Economics, is an expert in econometrics, the study of the methods used in economics. But the purpose of his specialty defies simple boundaries. After all, the point of refining research methods is to make applied studies better — and to better grasp their limits.
“There are many fields in economics that answer socially significant questions,” Andrews says. “There are things it would be good for us to understand, but I often find myself interested in how sure we are about them. To what extent do we know the things we think we know? To what extent is there more to know, based on the uncertainty and degree of confidence? These issues of uncertainty matter because the answers to the substantive questions matter.”
Andrews’ core contributions to economics very much involve uncertainty and confidence. He first became known for research on “weak identification,” settings where key variables do not yield much information about the issues being studied. And he has published notable work about the challenges of building economic models.
Andrews’ work also illuminates larger ideas about how we use data. His most recent published paper examines the “winner’s curse” in the social sciences — the idea that programs testing well one time are too often chosen for implementation, when sometimes they performed well purely by chance, and are likely to perform worse the next time they are tried out.
Another major Andrews paper, from 2019, analyzed how much publication bias exists in academic journals — which can lean toward publishing dramatic findings rather than equally valid null results, as replication studies, in part, can reveal.
At times, Andrews’ work seems like the social science equivalent of an X-ray machine: He scans studies to look for problems under the surface. But Andrews does not only look for problems; he develops techniques to prevent them in the first place. In typical Andrews fashion, his papers on the winner’s curse and on replication bias both offer new methods for avoiding these pitfalls.
“It’s important to work on these tools because the tools are going to be used on important things,” Andrews says. “If you have a beautiful tool and it’s never used, is it a tool, or a work of art?”
Andrews is recognized as a leading-edge practitioner in his field. In 2021, he was given the John Bates Clark Medal, awarded annually by the American Economic Association to the best economist under the age of 40. In 2020, he was granted a prestigious MacArthur Fellowship. With his career flourishing, Andrews rejoined the MIT faculty last year.
Career-changing conversation
Andrews grew up in the Boston area, in a family where both of his parents had earned PhDs in economics. While Andrews was not always set on becoming an economist, he did take advanced courses in the subject as an undergraduate at Yale University, where he graduated summa cum laude. He then entered the doctoral program at MIT.
At the time, in 2009, after the financial-sector meltdown and related recession, a lot of attention in economics was directed toward finance and macroeconomics, but Andrews did not feel compelled to study those topics.
“I didn’t feel they were such a good fit in terms of the style of work that appealed to me,” Andrews says. “I was finding econometrics-y questions very interesting.”
At one MIT Department of Economics function, Andrews started talking to Anna Mikusheva, an econometrician on the Institute faculty. By the end of the event, Mikusheva had suggested Andrews help with some research she was working on.
“The research assistant role for Anna turned into a joint project and I found my interest continuing to be drawn to these questions,” Andrews says. “So by virtue of that, that’s where my work went.”
Mikusheva and Andrews co-authored a high-profile series of papers on weak identification that wound up getting published soon after he received his PhD from the Institute in 2014. After spending a couple of years as a junior fellow in the Harvard Society of Fellows, Andrews joined the MIT faculty in 2016. He moved to Harvard University in 2018, then returned to MIT last summer.
As Andrews’ career has evolved, his wide-ranging work has often involved productive research partnerships, including papers co-authored Mikusheva, Matthew Gentzkow and Jesse M. Shapiro, Toru Kitagawa and Adam McCloskey, and Maximilian Kasy.
At all turns, Andrews stays focused on questions about the certainty (or uncertainty) involved in economic analysis and the degree of confidence (or lack thereof) we might have as a result.
“The worst scenario is when the data tells us very little, but we’re wrongly overconfident and think the data is telling us a lot,” Andrews says.
Making numbers more useful
To a consequential degree, Andrews also finds motivation in particular research problems. His recent paper (with Kitagawa) and McCloskey on the “winner’s curse” both introduces a new technique for estimating results and then applies it to a major research project on the social mobility of different U.S. neighborhoods, initiated by economists Raj Chetty and Nathaniel Hendren. Andrews’ conclusion: The Chetty/Hendren findings hold up well, suggesting social programs can productively use those results.
“The thing that matters for society is what is it we all do with those numbers,” Andrews says. “If we can think about what makes numbers more useful for people downstream, it’s important.”
Research is hardly all Andrews spends his time on. He has a long-running commitment to teaching, working with both undergraduates and graduate students, and as a PhD candidate in 2014, won MIT’s Robert M. Solow Prize for Excellence in Reasearch and Teaching.
“MIT students are very smart, so if you can help them frame a question the right way, it’s important,” Andrews says. “Ten years down the road, they may not retain the exact answer, but if they retain the framing of the question, they can work their way to the correct answer.”
In the MIT Department of Economics, where working productively with graduate students is a point of emphasis, Andrews now finds himself in the position Mikusheva was in, a decade ago, when she was encouraging him to follow his core intellectual interests.
“The graduate mentoring piece is very, very important,” Andrews says. “If you look at the social impact of an hour of my time, I feel the highest marginal product things I do are around advising. These are extremely capable people where a little bit of input or redirect can have big benefits down the road for them, and then hopefully the things they are doing are useful, and will benefit society.”
On all fronts, then, Andrews keeps trying to refine our knowledge about the extent of our knowledge. Summing up his work, Andrews offers his own epigram about the nature of his research.
“I would like to understand the extent to which we understand things,” Andrews says.
